Frontotemporal lobar degeneration (FTLD) is the second most common cause of dementia among subjects under the age of 65. There are two main FTLD subtypes: FTLD associated with lesions immunoreactive to TDP-43 (FTLD-TDP), and FTLD associated with Tau-immunoreactive lesions (FTLD-Tau). Currently, the exact FTLD pathology cannot be confidently defined until autopsy, and there is an urgent need for ante- mortem biomarker which reliably predict FTLD pathology. Over the past 30 months, I have identified and validated a panel of cerebrospinal fluid (CSF) FTLD-TDP biomarkers that identified FTLD-TDP patients with 85% accuracy, including Tau phosphorylation, tripeptidyl peptidase 1 (TPP1) levels, and inflammatory protein levels (FAS, eotaxin-3, IL-23), with 85% accuracy in distinguishing between the two main FTLD subtypes. The identification and validation of these CSF alterations will serve as the basis of the current proposal to translate these markers towards clinical application through multi-site technical validation and advanced classification through machine-based learning, and to better characterize altered biochemical pathways in FTLD-TDP. First, I hypothesize that we can reduce the coefficient of variation for each FTLD-TDP biomarker to under 15% through a multi-site technical validation using 60 banked CSF samples. I will test this hypothesis in Aim 1 by determining technical factors (freeze-thawing, blood contamination, detergent use, protease inhibitor use, buffer condition) which reduce assay precision, and validating the assay at Penn. Second, I hypothesize that machine-based learning/support vector machine approach will better diagnose FTLD-TDP by taking into account non-linear effects of age, gender, and disease duration. I will test this hypothesis in Aim 2 by comparing the classification performance between more established algorithms and the support vector machine algorithm. Lastly, I hypothesize that these biomarker changes reflect altered biochemical pathways in FTLD-TDP. I will test this hypothesis by measuring brain and CSF levels of proteins involved in Tau phosphorylation, TPP1 maturation, and inflammation. As an exploratory Sub-aim 3a, I will determine if any of the changes from Aims 1 & 3 are detectable in asymptomatic subjects carrying familial FTLD-TDP mutations, to power an R01 application on longitudinal CSF biomarker changes in prodromal FTLD-TDP subjects. I will carry out this proposal with guidance from my mentoring team including Allan Levey, MD, PhD (Emory), John Trojanowski, MD, PhD (Penn), James Lah, MD, PhD (Emory), Jonathan Glass, MD (Emory), Leslie Shaw, PhD (Penn), and Eva Lee, PhD (Georgia Tech). I will also obtain formal training in biostatistics, bioinformatics, analytical chemistry, quality control, clinical trial design, geriatrics, and responsible conduct of research. Successful completion of the aims in the current proposal will establish the standard procedures to clinically translate promising FTLD-TDP biomarkers, determine the best algorithm to diagnose FTLD-TDP using these biomarkers, and identify altered CSF and brain pathways related to these biomarker alterations.